When originally planned, the International Satellite Cloud Climatology Project (ISCCP) was to be a 5-year project, maybe lasting a little longer. Absent any other way to continue collecting cloud observations, ISCCP was extended and has now lasted nearly 30 years. From the beginning, the small staff available required more computer automation of the processing than usual for a scientific data analysis project. As the time record increased in length, it became more necessary to perform additional tasks like calibration and quality checking (QC) that had not originally been envisioned. Consequently, the current system grew in stages and not necessarily in an efficient way. Now, however, the ISCCP system has been re-engineered to make the design more consistent and efficient. The ISCCP data processing system is notable for several features: multi-agency (multi-national) data exchange arrangements and cooperative processing, common absolute calibration of all VIS-IR radiances from the whole constellation of operational weather satellites, a multi-stream into single-stream analysis procedure, and a hierarchical product design. Lessons learned from the development of the ISCCP processing system and from its recent conversion from a research-based to a fully operational data processing system will be discussed, including why multi-level computer automation was required, why QC had to performed at many stages of the processing, why the processing chain had to have multiple re-entry points, why re-processing will always be required to achieve climate data record quality, and why such a processing system will always require expert human involvement.